AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

TextPecker: Rewarding Structural Anomaly Quantification for Enhancing Visual Text Rendering

Hanshen Zhu, Yuliang Liu, Xuecheng Wu, An-Lan Wang, Hao Feng, Dingkang Yang, Chao Feng, Can Huang, Jingqun Tang, Xiang Bai

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Personalized and region-controlled generation · 13 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add TextPecker structure-aware reward in RLFlux.1[dev] · GenTextEval EnglishSem.: 0.336 -> 0.719Improved
use TextPecker reward instead of OCR rewardFlux.1[dev] · GenTextEval EnglishSem.: 0.602 -> 0.719Improved
use TextPecker reward instead of OCR rewardSD3.5-M · CVTG-2K EnglishSem.: 0.671 -> 0.593Hurt
use TextPecker reward instead of OCR reward on ChineseQwen-Image · GenTextEval ChineseSem.: 0.874 -> 0.897Improved
add TextPecker reward in RLQwen-Image · OneIG EnglishAvg.: 0.871 -> 0.990Improved
add synthesized data to annotated data (ablation)TextPecker (InternVL3-8B) · TSAP ChineseF1: 0.727 -> 0.927Improved
add synthesized data to annotated data (ablation)TextPecker (InternVL3-8B) · CTR Chineserecall: 0.849 -> 0.962Improved
train on annotated structural data versus base model (ablation)InternVL3-8B · TSAP EnglishF1: 0.183 -> 0.874Improved
add pairwise matching to OCR reward (ablation)Flux.1[dev] · GenTextEval-ENSem.: 0.602 -> 0.644Improved
add pairwise matching to OCR reward (ablation)Flux.1[dev] · GenTextEval-ENQua.: 0.976 -> 0.976No clear effect
replace OCR model with TextPecker in reward (ablation)Flux.1[dev] · GenTextEval-ENSem.: 0.644 -> 0.702Improved
add structural quality (SQ) reward (ablation)Flux.1[dev] · GenTextEval-ENSem.: 0.702 -> 0.719Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy